Make a strict specification and 1.0 release

[NikVolf]

Since trie logic is now part of the consensus, to avoid future reimplementation of similar prs (#42), implementation should follow strict specification and then never change in observable way.

[burdges]

Any chance we could get someone to review if this design is optimal first? Ala the two first comments in #23

I hear Vitalik considers non-fixing suboptimal trie designs to be among their larges mistakes with Ethereum.

[NikVolf]

Sure! We can start dedicated topic somewhere in research repos

[burdges]

I suppose this crate eventually becomes trie1 or whatever, and then we've do another next generation trie2 somewhere, and maybe another with a zk snark friendly hash.

I think migration paths exits: We eventually want "polkadot" to be a parachain of the polkadot relay chain, so maybe this future separate relay chain could launch with the new trie, after first disabling XCMP on the polkadot chain, or something like that.